May 23, 2023

Palette 3.4 brings a dozen new features, from edge K8s to security and observability

Anton Smith
Anton Smith
Director of Product

Palette 3.4 is here, and it’s more than just a point release. It’s jam-packed with enhancements and new features — driven by your feedback.

There’s something here for you, whether you're managing Kubernetes on edge hosts, deploying in a multi-cloud environment, or harnessing the power of GPUs for AI applications.

Read on to get the lowdown on a dozen of our favorite new features, or jump straight to:

Building a better edge

With Palette 3.4, we've been focusing on enhancing our edge capabilities. Let's dive in.

Edge host debugging: your early warning system

Deploying new edge hosts is a high stakes activity. Networking, hardware and power issues can all occur before a cluster is formed, and without detailed information on these problems you’re flying blind — you may even need to incur the cost and time of a site visit to achieve a fix.

With our latest release, every event from a remote edge host is logged, even before it becomes part of a cluster. You can now start troubleshooting and debugging edge hosts as soon as they boot up, reducing site visits and speeding up problem resolution.

edge host debugging

Static IP support: stability in a dynamic world

It’s a common challenge that you may need to deploy Kubernetes clusters in the field, including in DHCP network environments. You need to be sure that your clusters will be robust and continue to function even if their IP addresses change.

Now Palette 3.4 introduces the ability to assign static IP addresses to edge hosts during configuration, providing much-needed stability when deploying nodes in existing DHCP network environments. Other networking details, like DNS, netmask and routes, can still be inferred from the DHCP server.

static IP support - stability in a dynamic world

Customizable product UID: grab identifiers from any hardware

Every hardware device has its own unique way of displaying identifiers such as serial numbers, and sometimes you might need to follow specific organizational policies for what and how you capture this data. Palette 3.4 brings a flexible solution to this challenge, providing an easy way to extract a product device identifier from wherever you like.

Simplicity and power for IT operations

Our new features for IT operations teams are tailored to give you increased control, enhance security, and provide deeper insights into your Kubernetes clusters.

Observability stack: actionable insights, out of the box

Our new release introduces out-of-the-box monitoring and observability packs based on Prometheus for monitoring and alerting, and Grafana for visualizing the data.

These packs go beyond just reporting on the state of your clusters. They provide key operational metrics, performance data, and resource utilization, transforming raw data into actionable insights.

Palette supports flexible deployment options for these observability features, with both local and remote monitoring. With Palette it’s easy to deploy a monitoring stack along with all necessary components in every deployed cluster to report metrics back.

observability stack

Look out for a blog coming soon going deeper into the observability stack.

RHEL support on AWS: bring your own OS choice

Choice is essential when it comes to your operating system. Whether it's due to specific security, compliance, or hardware configuration needs, or simply because you want to offer your developers more options — Palette 3.4 has got your back.

We’ve introduced support for Red Hat Enterprise Linux (RHEL) on AWS. But we're not just stopping at adding another OS to the list. Our BYOOS (Bring Your Own Operating System) pack allows you to easily upload your own OS images, configure the necessary drivers, and customize the OS to meet your specific requirements. So if you need a customized RHEL environment, you've got it.

profile layers

EKS Launch Templates: more control, less hassle

You asked for it, we delivered. With the introduction of EKS Launch Templates support in Palette 3.4, you can specify what runs on your worker nodes right from the Palette platform. There's no need to hop between different tools or interfaces. Moreover, this customization can be incorporated into your cluster profiles. This means you can apply them across all your clusters, ensuring uniformity.

How it works: EKS Launch Templates

Here's a quick rundown of how you can use EKS Launch Templates:

  1. Create a custom AMI image and store it on AWS.
  2. Reference the AMI from the K8s layer in the cluster profile.
  3. Specify additional parameters such as instance type, root volume, and ssh key pair.
eks launch template

IRSA support: centralized management for AWS infra clusters

Many organizations are looking to centralize management of credentials in AWS, using its IAM (Identity and Access Management) and IRSA (IAM Role for Service Accounts) capabilities to avoid creating, injecting, and managing credentials directly inside the Kubernetes clusters — and the associated risk of leakages and outages resulting from accidents with credential management.

IRSA is complex and difficult to configure, but Palette makes it easy. Palette automates the setup of IRSA on AWS infrastructure clusters through simple declarative management, enabling users to get up and running quickly, while still providing for any customizations necessary.

Palette makes IRSA simple

Here's how Palette simplifies the IRSA setup:

  1. Palette automatically creates an OIDC provider and the necessary trust policies to enable IRSA.
  2. It also takes care of creating the necessary IAM Roles for your cluster and maps them to pre-existing or custom trust policies.
  3. With Palette, it's easy to apply the IAM Role mappings to applications across clusters at scale using pack macro expansions.

One major benefit you should know about: Palette’s support for AWS Infra enables the usage of IRSA for IaaS with custom CNI and CSI, which is not possible with EKS.

Custom registry switching with imageswap: simplifying multicloud image management

Dealing with multiple cloud providers and trying to configure them to use your custom container registries can be a chore. Whether it's for pull-through cache reasons or dealing with private container images, the process can be tedious and error-prone.

Palette 3.4 tackles this problem head-on with the image swap project. With image swap, you can easily match patterns and replace them, allowing you to change from one registry (like to another, effortlessly. Need to match exact paths including the full image name and rewrite it? No problem!

Palette's cluster profiles allow these modifications to be implemented at scale. That means less time wrangling registries and more time focusing on the work that matters.

Time-saving features for app developers

Palette 3.4 continues our journey to make developers' lives easier with new features for Palette Dev Engine (PDE). Forget about wrestling with Kubernetes clusters: let PDE handle the heavy lifting, so you can focus on your code.

PDE Dashboard: your operations at a glance

Do you want to quickly gauge your application performance, resource usage, or cluster health? Say hello to the PDE Dashboard.

It's a beautifully designed (if we do say so ourselves) hub that provides a quick overview of your running Palette Virtual Clusters and deployed applications. You can also monitor CPU, memory, and storage usage to optimize your resources and manage costs.

PDE Dashboard

Beyond the dashboard, we’ve also looked at how we can add more insights across the PDE user interface. This includes new pack information.

As your organization starts to use PDE more and more, you’ll find that you and your peers start to build up a library of packs that you assemble into your app profiles.

Naturally it’s important to know at a glance what you’re using, so in 3.4 we’ve enhanced the information you see in the user interface about each pack component, including the version and address of the registry hosting the pack. It should make your life a bit easier.

Object storage

PDE CLI: where devs feel at home

We understand that as a developer, your comfort zone is your terminal or IDE, and you want to spend as little time as possible jumping between different vendor portals and graphical UIs.

With the new PDE CLI, you can now manage your virtual clusters and PDE applications right from your local command line, without needing to leave your IDE.

It’s one more step towards reducing friction and wasted time, which has gotta be good for your metrics, right?

PDE applications right from your local command line

PDE on self-hosted Palette: more control, more security

While Palette is primarily a SaaS platform, did you know that many of our customers choose to self-host Palette in their own infrastructure, for example due to security or policy constraints?

Until now, Palette Dev Engine has only been available for organizations using Palette’s default SaaS delivery model. But as of Palette 3.4 we now support PDE on multi-tenant SaaS, single-tenant SaaS and self-hosted instances of Palette.

PDE on self-hosted Palette

Nvidia GPU support made easy

One of the most challenging aspects of running GPU-intensive applications on Kubernetes is getting the configuration right. From the operating system layer up to the application layer, there are a plethora of settings and drivers that need to be carefully aligned to fully harness the power of GPUs. It's a tricky process and can be quite a headache.

Nvidia GPU

With our new out-of-the-box Nvidia GPU Operator pack, we've taken the hassle out of configuring your clusters for GPU utilization.

The pack uses the operator framework within Kubernetes to automate the management of all Nvidia software components needed to provision a GPU.

These components include the Nvidia drivers (to enable CUDA), Kubernetes device plugin for GPUs, the Nvidia Container Toolkit, automatic node labeling using GFD, DCGM based monitoring and others.

Our Nvidia GPU Operator pack ensures that every layer of your cluster, right down to the operating system, is correctly configured to make the most of your GPUs.

But it doesn't stop there. With Palette's ability to manage configurations at scale, not only do you have the assurance that your clusters are GPU-ready, but also the peace of mind that comes with knowing this readiness can be replicated across all your clusters.

Want to learn more about how people are using GPU accelerators at the edge? See it for yourself with this incredible presentation at Kubecon Europe 2023 from Spectro Cloud and Tevel Tech — there’s nothing cooler than flying robots picking apples!

Ready to take Palette for a spin?

We've covered a lot of ground here. From the robust edge host management, simplified observability, customizable operating system choices, to developer-friendly enhancements with PDE, and the power-packed Nvidia GPU pack… Palette 3.4 truly has something for everyone! For all the technical details, don’t forget to check out our docs portal for the full release notes.

We're thrilled to see Palette 3.4 in your hands, and we can't wait to see the amazing things you'll achieve with it. As always, we're here to support you on this journey, and we're already looking forward to your feedback to make Palette even better.

Ready to discover the transformative capabilities of Palette 3.4? Try Palette today absolutely free, and get up to speed fast with all the step-by-step tutorials at

Subscribe to our newsletter
By signing up, you agree with our Terms of Service and our Privacy Policy